Attention:The NSF Public Access Repository (PAR) system and access will be unavailable from 11:00 PM ET on Thursday, August 13 until 12:00 AM ET on Friday, August 14 due to maintenance. We apologize for the inconvenience.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Tiny Language Models Enriched with Multimodal Knowledge from Multiplex Networks
Large transformer language models trained exclusively on massive quantities of text are now the standard in NLP. In addition to the impractical amounts of data used to train them, they require enormous computational resources for training. Furthermore, they lack the rich array of sensory information available to humans, who can learn language with much less exposure to language. In this study, performed for submission in the BabyLM challenge, we show that we can improve a small transformer model’s data efficiency by enriching its embeddings by swapping the learned word embeddings from a tiny transformer model with vectors extracted from a custom multiplex network that encodes visual and sensorimotor information. Further, we use a custom variation of the ELECTRA model that contains less than 7 million parameters and can be trained end-to-end using a single GPU. Our experiments show that models using these embeddings outperform equivalent models when pretrained with only the small BabyLM dataset, containing only 10 million words of text, on a variety of natural language understanding tasks from the GLUE and SuperGLUE benchmarks and a variation of the BLiMP task.  more » « less
Award ID(s):
2140642
PAR ID:
10684157
Author(s) / Creator(s):
; ; ; ;
Publisher / Repository:
Association for Computational Linguistics
Date Published:
Page Range / eLocation ID:
19 to 29
Format(s):
Medium: X
Location:
Singapore
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; (, KDD'21:The 27th {ACM} {SIGKDD} Conference on Knowledge Discovery and Data Mining, August 14-18, 2021)
    null (Ed.)
    Recent years have witnessed the enormous success of text representation learning in a wide range of text mining tasks. Earlier word embedding learning approaches represent words as fixed low-dimensional vectors to capture their semantics. The word embeddings so learned are used as the input features of task-specific models. Recently, pre-trained language models (PLMs), which learn universal language representations via pre-training Transformer-based neural models on large-scale text corpora, have revolutionized the natural language processing (NLP) field. Such pre-trained representations encode generic linguistic features that can be transferred to almost any text-related applications. PLMs outperform previous task-specific models in many applications as they only need to be fine-tuned on the target corpus instead of being trained from scratch. In this tutorial, we introduce recent advances in pre-trained text embeddings and language models, as well as their applications to a wide range of text mining tasks. Specifically, we first overview a set of recently developed self-supervised and weakly-supervised text embedding methods and pre-trained language models that serve as the fundamentals for downstream tasks. We then present several new methods based on pre-trained text embeddings and language models for various text mining applications such as topic discovery and text classification. We focus on methods that are weakly-supervised, domain-independent, language-agnostic, effective and scalable for mining and discovering structured knowledge from large-scale text corpora. Finally, we demonstrate with real world datasets how pre-trained text representations help mitigate the human annotation burden and facilitate automatic, accurate and efficient text analyses. 
    more » « less
  2. ; (, Findings of the Association for Computational Linguistics: EMNLP 2020)
    null (Ed.)
    In standard methodology for natural language processing, entities in text are typically embedded in dense vector spaces with pre-trained models. The embeddings produced this way are effective when fed into downstream models, but they require end-task fine-tuning and are fundamentally difficult to interpret. In this paper, we present an approach to creating entity representations that are human readable and achieve high performance on entity-related tasks out of the box. Our representations are vectors whose values correspond to posterior probabilities over fine-grained entity types, indicating the confidence of a typing model’s decision that the entity belongs to the corresponding type. We obtain these representations using a fine-grained entity typing model, trained either on supervised ultra-fine entity typing data (Choi et al., 2018) or distantly-supervised examples from Wikipedia. On entity probing tasks involving recognizing entity identity, our embeddings used in parameter-free downstream models achieve competitive performance with ELMo- and BERT-based embeddings in trained models. We also show that it is possible to reduce the size of our type set in a learning-based way for particular domains. Finally, we show that these embeddings can be post-hoc modified through a small number of rules to incorporate domain knowledge and improve performance. 
    more » « less
  3. ; (, Association for Computational Linguistics)
    Large scale transformer models, trained with massive datasets have become the standard in natural language processing. The huge size of most transformers make research with these models impossible for those with limited computational resources. Additionally, the enormous pretraining data requirements of transformers exclude pretraining them with many smaller datasets that might provide enlightening results. In this study, we show that transformers can be significantly reduced in size, with as few as 5.7 million parameters, and still retain most of their downstream capability. Further we show that transformer models can retain comparable results when trained on human-scale datasets, as few as 5 million words of pretraining data. Overall, the results of our study suggest transformers function well as compact, data efficient language models and that complex model compression methods, such as model distillation are not necessarily superior to pretraining reduced size transformer models from scratch. 
    more » « less
  4. ; ; ; (, IEEE)
    Humans excel at efficiently navigating through crowds without collision by focusing on specific visual regions relevant to navigation. However, most robotic visual navigation methods rely on deep learning models pre-trained on vision tasks, which prioritize salient objects—not necessarily relevant to navigation and potentially misleading. Alternative approaches train specialized navigation models from scratch, requiring significant computation. On the other hand, self-supervised learning has revolutionized computer vision and natural language processing, but its application to robotic navigation remains underexplored due to the difficulty of defining effective self-supervision signals. Motivated by these observations, in this work, we propose a Self-Supervised Vision-Action Model for Visual Navigation Pre-Training (VANP). Instead of detecting salient objects that are beneficial for tasks such as classification or detection, VANP learns to focus only on specific visual regions that are relevant to the navigation task. To achieve this, VANP uses a history of visual observations, future actions, and a goal image for self-supervision, and embeds them using two small Transformer Encoders. Then, VANP maximizes the information between the embeddings by using a mutual information maximization objective function. We demonstrate that most VANP-extracted features match with human navigation intuition. VANP achieves comparable performance as models learned end-to-end with half the training time and models trained on a large-scale, fully supervised dataset, i.e., ImageNet, with only 0.08% data. 
    more » « less
  5. ; ; ; (, International Conference on Learning Representations (ICLR 2026))
    Modern large language models (LLMs) excel at tasks that require storing and retrieving knowledge, such as factual recall and question answering. Transformers are central to this capability because they can encode information during training and retrieve it at inference. Existing theoretical analyses typically study transformers under idealized assumptions such as infinite data or orthogonal embeddings. In realistic settings, however, models are trained on finite datasets with non-orthogonal (random) embeddings. We address this gap by analyzing a single-layer transformer with random embeddings trained with (empirical) gradient descent on a simple token-retrieval task, where the model must identify an informative token within a length-L sequence and learn a one-to-one mapping from tokens to labels. Our analysis tracks the ``early phase'' of gradient descent and yields explicit formulas for the model's storage capacity -- revealing a multiplicative dependence between sample size N, embedding dimension d, and sequence length L. We validate these scalings numerically and further complement them with a lower bound for the underlying statistical problem, demonstrating that this multiplicative scaling is intrinsic under non-orthogonal embeddings. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Text Encoded Conversion
  • XML
  • Save / Share this Record
  • Send to Email